BFINTBFIVZATION W SOYBTAN OIL BY **BB OF GBNTBIFUGB

by

William D. Feagin

A Thesis Submitted for Partial

Fulfillment for the Degree of° BAGHBLGR nv scxmmcw

in

CHW"IOAL TWGIN?W°ING

Approved:

n charge of Major wer

Hea of e jor Department

4Dean of Tngineering Division

Virginia Polytechnic Institute

I

way 13. 1939

98957

gg G¤NT7NTS

PageI. Introduction.......................... 1

II. Ieview of Literature.................. 2

III. Wxperimental.......................... 5

A. Purpoee of Study................. 5

B. Plan of Inveetigat1on............ 5

C. ”ater1ale........................ 6

n, apparatus........................ 7

7. Yethod of Procedure.............. Il

V. Iabulated Wata and Tesults....... 12

IV. Wiecuseion............................ 15

V, Conclue10ns........................... 13

VT. Tummnry............................... 20

VII, Literature Cited...................... 21

VIII. äcknowledgemente...................... 22

Tables

Table I .................................. 14

rabxe xx ................................. 15

,1-

I. INTRODUGTION

The United States ls second only to Manchuria

in the production of soybeans. One of the most

strlking developments in the United States in re-

cent times is the rapid rise of the production of

soybeans from three million bushels in 1907 to

nearly forty million bushels in 1920. This large

yield necessitated outlets other than as stock feed

and so attracted the attention of the oil mllls,

which by 1929 began to be a major factor in the pro·

ductlon of the crop. In 1926 2,500,000 pounds of

soybean oil were produced, but in 1935 the crop of

oil was 200,000,000 pounds. There are many uses of

the oil, malnly as foodstuffs as a butter substitute,

cooking oil, lard substltute, salad oil, and as a

medicinal oil. The chief dlfficulty in the use of

this oil lies in the fact that lt becomes rancld soon

after production.

This problem deals with the eliminatlon of the

various components of the soybean oil, causing the

rancidity, by use of the supereentrifuge.

-2-

II. REVIEW gg LITERATURE

ßoybean oil ie very easily refined and deodo•

rized, but on storage it rapidly turns rancid, ao-

quiring a flavor described as grassy. Lewkowltsch

ascribes the rancidity to the formation of free

fatty acids in the preaence of moisture, with en-

zymee acting as accelerators for the reaction. In

crude oil the free fatty acids range around 0.5%.

indioating conslderable natural protection against

hydrolysis in the sound beans. Patty acids are cata-

lysts for the auto·exidat1on of glycerides, greater

molecular weight giving greater effect. The com-

position of the oil 16 as fcllowe:

GLYGERIEES ggg_Qggg

Linolenic 2,3

Llnolelc 51,5

Oleic 33.4

Palmitlc 6.8

Stearic 4.4

Arachidic 0.7

Lignoceric 0.l

Unsaponifiable 0.6

I

-3-

Nearly 9Q% of the fatty aclds present are unsaturated

with an iodine number of about 13ä.

The refining of the oil is usually carried out

ln five steps: neutralization, washlng, bleaching,

winterlzatlon, and deodorlzation. The only part of

this process with which we are concerned though is

the neutrallzatlon. Neutrallzation, as used here,

applles to the separation of the soap stock from the

neutral oil, the soap stock including free fatty

acids, lye, and nearly all lipolds present. In the

ordinary process this sepsratlon 1s carried out by

first emulsifying the oil at 709 F with l0•l49 Be.

sodium hydroxlde, and then on continued heat to

around l60° F, the soap clots form and separate out

after settling for l8•20 hours. Thls process ls also

carried out in a continuous method by the use of

first a small bester and then neutralizstlon by use

of a high speed centrifugal machine. The oil in

either oase should be entirely free of lipoids after

neutrallzation, the abscence being shown by the "Heat·

break" test, which consistz of freedom of soap stock

in the form of floc when a sample of the oil is heated

tc 600° r. (2).

The use of the centrifuge is also carrled on in

-4-

the removal of phospholipins from the oil. Phospho-

llpins are substltuted fats in which one molecule of

fatty acid is replacsd by a molsoule of phosphorlc

acid united with a nitrogen base. Two of these

phospholiplns -lecithln and cephalln are found in

soybesn oil and are removed commercially by treating

with steam and then C6hbP1fUg1Hß• The same result

ls also accompllsed by emulslfication with water

and oentrifuglng, which also gives a good grade of

non—break oil and orude phospholipins. (4).

15Q

III.A.

Purpess of Ttudy

The purpose of this study usa te determine the

ef°ect of centriduging upon soybean oil ss to

lessening ot rancidity end ixnrevement of taste

and quality.

B. Plsn of Fnvestigetien

1. Secure samples ef seybean oil.

2. Ymulsify samples end run through centrifwge.

verging the speed and the size ef ring dem used.

3. übserve and record the following date on each

sample:

s. lpeed

b. dize of ring dem used

o. @1me of run

d. dize of sample

6. ämcunt of oil rscevered

1

-5-

f. Appearance of oil immediately after

recovery

5. Appearance of 011 after eettling for

several days _

h. Odor

1. Taste

4. Hake "Heat·break" test en each sample of oil

recovered

C. Mater1als

1. Govbean gl; - The oil used was pressed from

beans of the TOKIO (glyclne hispia) variety,

introduced from Tokio, Japan, 1901. Plants

stout, erect, maturing in about 145 days;

pubescence wray; flowers both purple and white,

70 to 75 days to flower, pods gray 40 to 50 mm

long, 10 to 12 mm wide, 6 to 7 mm thick, 2-3

seeded, shattering little; seed olive yellow,

7 to 8 mm long, 6 to 7 mm wide, 5 to 6 mm thick;

hilum pale; germ yellow; oil 18.4 per cent; about

134,000 to the bushel. (3).

2. §g&gg_· The water used for emulsiflcation

was ordinary tap water.

-7-

D, Apparatuß

The supercentrifuge used was constructed by

the Sharples Specialty Company, Type·4l-23-80-34,

Serial Number 331A3l2, Vaximum bowl speed of

50,000 revolutions per minute. The machine was

set up ae shown in figure 1. The bowl, 9 , has a

total weight of 3.13 pounds, and an inner length cf

elght inches. The bowl has an outside diameter of

1.875 inches and as inside diameter of 1.75 lnches.

The bowl rotates around its center of gravity and

is suspended from the turbfne 25 by means of a

flexible spindle 18 , This suspension of the

bowl enablee the machine to run with a minimum of

power and a minimum of wear. The bowl, when turning

at a speed of 40,000 revolutions per minute, gene·

rates a separating force of over forty-two

thousand times the force of gravity.

The centrifuge is so designed that it may be

operated to separate solide from liquids, or to

separate llquids of different specific gravity. The

method of operation is determined by the size of

ring dam 7 used. The ring dam is a small brass

washer which controls the amount of liquid allowed

T

l-3-

te enter the upper cover 22 . The outside die-

meter of the ring dame le 1.4375 lnches, but thß

diameter ot the inner hole varies from 0.75 to

1.0 inchea. The heavy liquid eacapee into the upper

cover while the llghter liquid eecapee into the

lower cover P3 . The liquid tc be centrituged

is fed to the bowl from a glass tunnel, connected

by a rubber hose to the feed noszle. The tunnel in

so arranged that a head of about two feet ls main-

tained by the liquid to be fed ln. Three sizes of

feed nozzles were available. and used according tc

the rate of flow desired into the bowl.

A small glass bottle, holding about one quart,

was provided for use in emulsifylng the oil, and

other appropiate vescels were provided to collect

samples.

-11-

E. Method ef Procedure

QL; — The eil was already preesed from the beans,

and had acquird a rancid oder and grassy taste.

Samples of the eil were measured, and emulslfled

l by shaklng with water.·”

Centrifugg-

The centrifuge was etarted and the

speed gradually increased pntll the desired speed was

attained. The feed nozzle was then ineerted, and

water fed lnte the bowl until water hegen te es·

oape from the upper spout. The emulsion was then

slewly poured inte the funnel, with continued

shaklng so that the eil and water would het separate.

The llquide coming off at the upper and lower spouts

were caught in veeeels until ne more eil would come

from the lower speut. The machine was then cut off

and allowed to come to rest, and the liquid remaining

in the howl ceught in a vessel as the howl stopped

roteting. Samples were then taken ef the precipltate

in the bewl, and the taste and oder ef the oil noted

and recerded. The machine was then cleaned from oil

fer the next run.

-12-

gggpgggg - Samples of the oil recovered from

the lower spout were then put under the "Heat-break"

test by heating to 6009 W in the electric oven. Other

samples of the oil were then taken and the above pro-

cedure repeated at varylng speeds ad with different

sizes of ring dame in place. Samples of the oil

obtalned under the various conditions were placed in

10 cc sample bottles and the bottles stored for

future tests as to odor, taste, and color.

F. Tabulated Data and Results

Table l shows the effect of speed and changes

in size of ring dame on the quality and quantlty

of oil recovered.

Table II shows a comparison of the oil

recovered under the various conditions.

A sample of oil recovered by use of the $7 l/2

ring dam was rerun through the centrifuge. The

separation on the second run was very poor as most

of the oil was recovered, but the recovered oil con-

tained a great deal of flocculent preclpltate.

-15•

IV,Precipitate.The cloadiness in the oil

immediately after recovery was mostly due to en·

trapped air, caught during the rapid whirling of

the oil in the bowl. This cleudiness disappeared

after several days of settling, and a sliqht

flocculent precipitate appeared in each of the

samples. This precipitate seemed to be of the

same nature as that recovered in the bowl, and

therefore probably is a small portion of the soap

stock that had been precipitated yet unremoved

during the rapid haseage of the oil through the

bowl, The color of the oil after setäling was some-

what lighter than the color ef the original oil,

Qägg. The oder of the oil was much improved

by centrifuging, although the taste of the oil did

not seem to differ greatly from the original taste.

The precipitate in the bowl in each case had a very

rancid oder.

Qggg gg Qgggh, A centrifuge of this type can

not be used on a continuous basis as the howl fills

-16-

completely with the precipitated soap stock and pre-

vents further feeding. The quality of the oil is

also very poor and the renoval of the soap stock

is very slight as the bowl becomes full. The centri-

fuge con be used very satisfactorily though on a batch

basis with the size of the howl as the limiting

factor,

fing Qgg_Elgg, The ring dam, 7 l/2, seemed to

be the correct one to use for this separation as a

greater percentage of the oil is reoovered, and the

precipitate in the bowl was entirely free from oil,

and less precipitation in samples of this separa-

tion than in the others. It does not seem pro-A

fitable to run the separated oil through the cen-

trifuge again as a very poor separation was ob-

tained, although there seemed to be quite a bit of

soap stock in the oil recovered. however, doubt

that all of the material precipitated in the sam-

ple was soap stock was caused by the fact that sam-

ples of the eil gave only a very slight precipl-

tate under the "Heat-break" test. The oder and

taste of the oil seemed to be no better after the

second run than before. Therefore, it seems that

no benefits are obtained by running the oil twice

-17-

through the centrifuge.

§pggQ„ Variations in speed below maximum seem

to have little effect upon the separation except

that the higher speeds seem to precipitate the

soap stock better and more quickly in the bowl.

Feed1n5„ The smallest nozzle was used to feed

the emulsion to the bowl, as a better separation

seemed to be effected when the oil was fed in

more slowly. Some d1f”iculty was experienced

though es the nozzle tended to clog at times. U

-13-

V, C¤NCLUSIONB

The results of this investigation indäeate

that:

1. The best separation of eoap stock from

the soybean oil is obtained by use of

the # 7 l/2 ring dam.

2. The best separation ef eoap stock from

the soybean oil is obtained at maximum

speed.

3. The rancid oder of the soybean oil is

eliminated to a large extent by the

centrifugation of the oil.

4. The maximum yield of soybean oil is

obtained by separation with the # 7 l/2

ring dam.

5. The grassy taste of the soybean oil

is only slightly improved by centri~

fugation.

6. The color of the eoybean oil is only

slightly changed by centrifugation.

7. No additional benefits are obtained by

„lg-

passing the soybean oil through the cen·

trifuge a second time.

8. A centrifuge of this type can not be run

on a oontinuoue basis, when a separntion

of this type is being oarried out.

9. It is definitely ooncluded, after cen-

trifugation that most of the ranc1d1ty

in soybean oil had been removed and that

the oil was much more edible than formerly.

JV

r

-90-

VI. SUHv^RY

Samples of soybean 011 were centrifuged in

e Sharples supercentrifuge at epeeds varying from

15,000 to 45,000 revolutions per minute and with

ring dem sizes of ¥ 7 to ä 9 in use in order to

determine the effect of centrifuging upon the

rancidity, oder, color, and taste ef the eil,

It wae found that most of the rancidity could

be removed, the taste improved slightly, the color

only slightly changed, and the oder of the eoybean

oil imprcved very much. In general, the iuallty of

the oil was made much more deeirable although thirty·

_ five per cent of the oil was lost during the prodese

es eoap stock.

VVVV c